Principal Software Architect

Oho Group; San Francisco (CA); On-site; USD 180,000 - 280,000; Full time

Benefits offered by this job

Equity

Competitive compensation

Job summary

Oho Group is seeking a Lead Architect to define how state-of-the-art foundation models execute on next-generation silicon. This role will lead the architectural roadmap across the software stack, bridging high-level ML model execution with low-level hardware capabilities.

You will collaborate with chip design teams to influence features based on compiler and runtime requirements, and optimize lowering pipelines, code generation, and kernel performance for massively parallel execution.

Qualifications

• Experience in AI compilers and accelerator architectures.
• Production-level knowledge of MLIR/LLVM and parallel computing models.
• Track record delivering software-hardware co-design solutions from concept to deployment.

Responsibilities

• Define the architectural roadmap across the software stack, bridging ML model execution with hardware.
• Collaborate with chip design teams to influence hardware features based on compiler/runtime needs.
• Optimize compiler lowering, code generation, and kernel performance for massively parallel execution.

Skills

AI compilers

Systems software

Accelerator architectures

MLIR/LLVM

Parallel computing models

Software-hardware co-design

Tools

MLIR

LLVM

Job description

We are representing a well-funded AI chip startup preparing to launch its flagship hardware platform into production.

They are looking for a Lead Architect to define how state-of-the-art foundation models execute on next-generation silicon.

What You’ll Do:

• Lead the architectural roadmap across the entire software stack, bridging high-level ML model execution with low-level hardware capabilities.
• Collaborate directly with chip design teams to influence hardware features based on compiler and runtime requirements.
• Optimize compiler lowering pipelines, code generation, and low-level kernel performance for massively parallel execution.

What We're Looking For:

• Deep expertise in AI compilers, systems software, or accelerator architectures.
• Production-level knowledge of modern compilation frameworks (e.g., MLIR/LLVM) and parallel computing models.
• Track record of bringing software-hardware co-design solutions from concept to deployment.

Join an experienced team at the point where architectural choices become customer-facing products, with competitive compensation and equity.

Keywords

Pytorch / ONNX / MLIR / LLVM / SIMT / CUDA / Kernel / Architect / GPU / Accelerator / ML / C++ / Workloads / Frameworks / AI